Our Methodology
Most comparison sites list what firms offer. PropFlux simulates what happens to you. This page documents every model, score, and number on the site — including the assumptions and the limitations.
FluxEngine — Pass Probability by Simulation
FluxEngine runs 10,000 Monte Carlo simulations of your trading — win rate, risk per trade, reward-to-risk ratio, trades per day — against each firm's exact evaluation rules. Not a rules summary. The actual mechanics: how the trailing floor updates, whether the daily loss limit fails you or just ends your day, how the consistency percentage is computed.
Each simulated evaluation runs day by day until it passes, breaches, or times out. Across 10,000 runs that yields, per firm and account size:
- Pass rate — the share of simulated evaluations that hit target without breaching a rule
- Median days to pass — and the full distribution, not just an average
- Failure breakdown — which specific rule ends the runs that fail (trailing drawdown vs daily limit vs time-out)
- Risk of ruin and estimated cost to fund — expected attempts times real fees
The same engine keeps running after the evaluation. The funded-account model simulates payout ladders, qualifying-day requirements, safety-net buffers, and funded consistency rules to estimate your probability of a first payout, survival at 3 and 6 months, and realistic income ranges.
We also tell you how sure we are. A win rate estimated from 50 trades carries real statistical error (standard error = √(p(1−p)/n)). We propagate that through the simulation by re-running it at the bounds of an 80% interval on your true win rate, and report the resulting pass-probability range alongside the point estimate. A "62% (52–71%)" result is honest; a bare "62.3%" would be false precision. The more trades your stats are based on, the tighter your range.
Bootstrap mode: your real trades, not assumptions. Paste your actual per-trade P&L and the engine stops assuming anything about your trading. Trades are normalised to R-multiples (average losing trade = 1R) and each simulated trade is resampled from your real distribution — preserving your actual fat tails, oversized losses, and home-run winners that a fixed win-rate model flattens away. Behavioural and market layers still apply, implemented as a shift in the resampling odds. This is standard bootstrap methodology, and it is the difference between simulating a trader and simulating you.
The 7-Layer FluxModel
A naive simulator that flips a weighted coin once per day would overstate everyone's pass rate. Real traders tilt, real trades have intraday excursion, and real markets cluster. FluxModel layers each of these effects explicitly — and you can toggle them on or off in FluxEngine to see exactly how much each one moves your numbers.
Behavioral Decay Function
Models psychology under pressure. Consecutive losing days, drawdown fear, and post-loss overconfidence all degrade the simulated win rate (capped at −25 points). By design, psychology can only hurt performance in the model — never help.
Intraday Variance Model
Simulates per-trade excursion (MFE/MAE) instead of treating each day as a single number. This is what catches intraday trailing drawdowns that punish open-profit peaks even on days that close green.
Firm Rule Engine
A unified breach engine encoding each firm's exact rules: trailing vs static vs EOD drawdown, daily loss limits (including cap-only DLLs that pause rather than fail), consistency thresholds, minimum trading days, and calendar time limits.
Time Decay Pressure
For firms with calendar-day evaluation windows, simulated traders take on more risk as the deadline approaches — matching the desperation behaviour we see in real failed evaluations.
Market Regime Simulation
Days are drawn in trending / normal / choppy regime blocks rather than independently, so win streaks and losing streaks cluster the way real markets make them cluster.
Funded Account Return Model
Runs the same engine past the evaluation: payout ladders, qualifying winning days, safety-net buffers, funded consistency rules, and payout frequency — producing survival curves and income distributions, not just a pass/fail.
Composite Index
A 0–100 index combining pass probability, cost efficiency, funded survival, income expectancy, and rule friction into one comparable number per firm and account size.
FluxDiagnose — Failure Post-Mortems
FluxDiagnose takes the daily P&L of a failed evaluation and classifies why it failed: a single catastrophic day, gradual erosion, or a consistency-rule trap. It scans for behavioural patterns — revenge trading after losing days, escalating trade counts, tilt sequences — and pinpoints the critical moment the evaluation was lost.
Then it does the thing no directory can: it replays your exact trading against every other firm's rules. If the same P&L would have passed under another firm's drawdown structure, that's not a trading problem — it's a firm-selection problem, and we tell you so.
Trust Score
Each firm's Trust Score (0–100) is a weighted composite of five components, scored editorially from documented evidence — payout records, published rule changes, and our own funded-account experience. They are not user-submitted star ratings, and they are never influenced by affiliate commission rates. Every firm page shows the full component breakdown.
Payout reliability
Documented payout history, denial reports, and processing consistency. Weighted most heavily.
Rule transparency
How clearly rules are published, and how often they change without notice.
Customer support
Response quality and how disputes are handled.
Platform stability
Data feed quality, outages, and execution reliability.
Value for money
Fees, resets, and activation costs relative to what the account actually offers.
True Cost to Funded & ROI
The advertised evaluation fee is almost never what getting funded actually costs. Our cost model uses your simulated pass rate:
- Expected attempts = 1 ÷ pass rate. A 40% pass rate means ~2.5 paid attempts on average.
- Total cost to funded = (discounted fee × expected attempts) + activation fee + first month of data/platform fees.
- Break-even = total cost ÷ expected monthly funded income, where income comes from the funded simulation and already accounts for the probability of blowing the funded account.
- Expected value = pass rate × expected 6-month funded income − total evaluation cost. Negative EV means the math says don't buy.
- Probability of profit = P(pass) × P(surviving funded long enough to recoup costs).
This is why our rankings sometimes disagree with "cheapest fee wins" lists: a $49 evaluation you fail five times costs more than a $150 one you pass twice — and a 90% profit split is worth little at a firm where the payout ladder caps your first six withdrawals.
Data Verification Policy
Prop firm rules change constantly — often without announcement. Our policy:
- Every firm profile carries a "rules verified" date, shown on the firm's card and page (most recent verification: July 5, 2026).
- Figures we haven't re-confirmed against the firm's current published rules are flagged internally and prioritised for re-verification — we'd rather show a flagged number than a confidently wrong one.
- Simulation parameters are updated whenever a firm changes drawdown mechanics, consistency thresholds, or payout structure.
Spot something outdated? Email hello@propflux.com — corrections ship fast and we credit reporters.
How We Make Money
PropFlux earns affiliate commissions when you sign up with a firm through our links. Two structural safeguards keep that from biasing results: rankings are produced by the simulation engine, which has no input for commission rates, and every formula is documented on this page, so you can check the math yourself. If a firm scores badly, it scores badly — affiliate or not.
Limitations — Read This Part
Honest models come with caveats. Ours:
- Your inputs drive everything. If you enter a 60% win rate you don't actually have, the pass probability is fiction. Use at least 30–50 real trades to estimate your stats.
- Simulated psychology is a model, not a measurement. The behavioural layers are calibrated to be directionally right, not personally precise.
- Slippage, commissions, and news events are folded into your win rate and R:R inputs rather than modelled separately.
- Past rules aren't future rules. A firm can change its drawdown mechanics tomorrow. Check the verification date.
Nothing on PropFlux is financial advice. It's a decision tool — the numbers a prop firm would rather you didn't calculate before buying an evaluation.